Re: [Monotone-devel] Re: Future of monotone

[Thomas Moschny]

On Monday 28 January 2008, Bruce Stephens wrote:
> Locality on disk, such that commonly the certs that you want to
> request are close together (and so more efficient to read).
>
> I guess lumping certs associated with a revision together might be
> good, but maybe not---maybe you want certs for a particular branch
> close together, and that that's much more important.

Well, I wasn't talking about performance, but about semantic problems, e.g. 
the problem that currently certs which are issued in one logical step are only 
loosely coupled later on; or that you don't have author or date information 
(or a message fwiw) attached to the 'tag', 'suspend' or 'test-result' certs.

- Thomas

signature.asc
Description: This is a digitally signed message part.